AC Transient Voltage and Xerox Copy Machines.

[CaptainObvious]

Let me preface this by saying this is for my personal curiosity and interest, even though it's business related.

I've been working with a Xerox company for a little over a year now, and we've had ongoing issues with some of the copy machines we sell to consumers, in particular with certain cities/villages in Alaska. The machines are generally 120v @ 15A. The machines experience various electrical issues, anything from sensors (such as a tray sensor, or jam sensor) will malfunction and will not work properly until the customer powers the machine down, then powers it back up. They'll work fine for an undetermined amount of time, then have another (or the same) issue come up.

The Xerox service repair technicians have been having us purchase "Power Filters" for these customers, which indeed seems to have resolved these issues.

Now, for my personal curiosity, I'm just curious if anyone here could provide some technical explanation on (potentially) what's happening and why the power filters seem to resolve these quirks? For a $10,000 to $100,000 copier.. (these are averages), I would have assumed they would have been able to handle this internally in the machine? Moar filtering maybe?

I would like to actually see the difference in "dirty power" and a "filtered power" from an outlet. Seems like the few videos that are prominent on YouTube are from quacks trying to sell books on "Toxic Radiation from your mains line!" and "Reducing EMF for your health", and I have a hard time taking them seriously.

[T3sl4co1l]

From what I've seen, the quality of electrical design is nearly independent of price.  90/10 rule and all that.

It may even be negatively correlated, since low cost, high production (millions qty/yr) stuff needs to be engineered to margins as razor-thin as the profit margins are, and they can amortize more NRE.

They have to send those machines through EMC testing (though if we're talking US market, they won't necessarily have been tested for susceptibility!), so they have to put up with some minimum amount of noise.  But that doesn't say anything about the behavior under higher noise levels -- whether it will pass, or fail softly, or need a full power cycle.

The symptom sounds like it might be one of the many cable harnesses inside, with poor RF grounding (there is a lot of plastic used in the construction of those machines), and susceptible logic inputs, say for opto-interrupters.

It's my personal policy to put RFI filtering and ESD protection on all pins, but it's also very easy for a company to hire an engineer less aware of, or concerned with, robustness, and you wind up with a huge expensive design, that electrically is little more than a fuckoff-massive Arduino shield with wires coming off in all directions.

Tim

[SeanB]

I would guess the AC power in the Alaskan areas comes from large diesel generators, running in a bank with the common bus bar being the local power feed. With this I would suspect the local grid there has a signifigant impedance, certainly more than the few milliohms of a larger grid, and more like in the hundreds of milliohms range. Thus any transient switching will easily propagate through the system, and can cause either high voltage spikes, or undervoltage events, which then lock up the microcontroller in the copier on random inputs to cause this.

The line filter likely is a large common mode choke, a lot of filtering class X capacitors ( probably 4u7 or larger) and some surge clamping as well. Adding this expensive part to all the machines would have been cost cutted in most cases as overkill, even though the original Xerox machines from decades ago would have had them as standard, so they could be used literally anywhere without failure. As the filter probably costs $50 or more to make they leave them out, relying instead on lesser amounts of filtering to handle the more robust supply area needs that the vast number of machines are used in.

I remember stripping old AM International ( remember them anybody) copiers, and, aside from having close to 200kg of steel plate internally as chassis parts, they also had a 30kg transformer in the base that provided both universal voltage compatibility ( the machine AC side was all 115VAC 50/60Hz motors and heaters, with speed selection done with pulleys that had 2 belt diameters and a guide that you moved to either a 50Hz or 60Hz position in the machine) along with doing voltage isolation as well, and giving a multitap selection from 90VAC to 255VAC as primary side taps on a tag board. Then you had another 200kg of moving parts as the copier mechanism. With maintenance those machines would easily run dozens of millions of page cycles, the only thing that killed them was the cost of maintenance, I know of a few that went over the 50 million page mark with care, as they were in use almost 24/7 in a university library.